🎵 Fonetizer

A tool that converts English song lyrics into singing-phonetic tables for vocal performance, optimized for vowel carrying, legato, and minimal consonant friction.

Perfect for barbershop quartets! 🎶

---

📚 User Guides

New to Fonetizer? Start here:

• 🇸🇪 Svensk guide (GUIDE_SV.md) - Komplett manual på svenska
• 🇬🇧 English guide (GUIDE_EN.md) - Complete manual in English

These guides include:

• Step-by-step instructions
• Understanding the phonetic table
• Tips for singers and quartets
• Frequently asked questions
• Technical details

---

Purpose

Fonetizer takes English song lyrics (one phrase per line with measure markers) and generates a phonetic table where each phrase is represented by one column pair:

• Left column = consonant(s), right-aligned
• Right column = vowel, left-aligned

These two columns are always read together for singing purposes.

Installation

Requires Python 3.11+ and dependencies.

# Clone or download this repository
cd Fonetizer

# Install dependencies
pip install -r requirements.txt

Usage

Option 1: Web App (Recommended for mobile) 📱

Streamlit Web Interface:

Run locally:

streamlit run app.py

Or deploy to Streamlit Cloud for mobile access! See DEPLOY.md for instructions.

Features:

• 🎵 Mobile-friendly interface
• ✏️ Paste and edit text directly
• 📝 Add line breaks and measure numbers
• 📥 Download formatted Excel file
• 🌐 Share with your quartet via link!

---

Option 2: Command Line (Desktop)

Generate formatted Excel file:

python fonetizer.py input.txt output.xlsx

This creates a fully formatted Excel file with:

• ✅ Borders around column pairs
• ✅ Right-aligned consonants
• ✅ Left-aligned, bold vowels
• ✅ Auto-adjusted column widths

Just open the .xlsx file - no manual formatting needed!

Generate CSV file:

# To file
python fonetizer.py input.txt output.csv

# To stdout
python fonetizer.py input.txt

# From stdin
cat lyrics.txt | python fonetizer.py

CSV files need manual formatting in Excel/Sheets.

Input Format

Each line can contain:

With measure numbers:

N <text>

Without measure numbers:

<text>

Where:

• N = starting measure number (optional, just the number, no "T" prefix)
• Space after the number
• <text> = the lyric phrase

Example with measure numbers:

1 There's a party tonight everyone is coming,
3 the music is high all the folks
5 enjoyin' and dancing the night away.

Example without measure numbers:

There's a party tonight everyone is coming,
the music is high all the folks
enjoyin' and dancing the night away.

Important:

• Exactly one phrase per line
• Measure numbers are optional
• If using measure numbers, only starting measure needed (no end measure)

Output Format

The output is a transposed table in CSV/TSV format:

• Rows = syllable positions (1, 2, 3...)
• Columns = phrase column pairs (consonant + vowel per phrase)

Column headers show measure intervals (T1, T4, etc.)

Example structure:

Syllable    T1      T4      T4      T8      T8      T10
1           aɪ      j       w       ɛ       ð       ɛ
2           j       uː      d       n       ɛ       n
...

Each phrase gets two columns:

1. Consonant cluster (right-aligned)
2. Vowel (left-aligned)

Phonetic Rules

Syllable Division

Each syllable is split as:

[CONSONANT CLUSTER] [VOWEL]

• Non-phrase-final consonants move to the next syllable
• Phrase-final consonants stay with their syllable

Diphthongs

All diphthongs are divided into two separate rows:

[Vː | glide]

Examples:

• cry → aː (row 1) + ɪ (row 2)
• know → oː (row 1) + ʊ (row 2)
• way → eː (row 1) + ɪ (row 2)

The second row (glide) has an empty consonant column.

Specific Pronunciations

• to → always uː
• to- in "tomorrow" → uː

Current Status

✅ Fully Functional! The current version implements:

• Full IPA phonetic transcription using the eng-to-ipa library
• Proper diphthong splitting into two rows
• Consonant cluster handling
• English-specific pronunciation rules from a comprehensive dictionary

The tool is ready for use with English song lyrics!

Possible Improvements

Future enhancements could include:

1. Fine-tuning syllable boundary detection
2. Adding support for singing-specific pronunciation variants
3. Handling multi-word phrases more intelligently
4. Export directly to Excel with formatting (bold vowels, cell borders)
5. Web interface for easier use

Example

Input:

T1-T4: I used to be the first one to cry

Expected output (when fully implemented):

Syllable    T1    T4
1           aɪ    j
2                 ɪ
3           j     uː
4           ...   ...

Contributing

This is a work in progress. The phonetic rules need to be fully implemented based on the specifications in CLAUDE.md.

License

Open source - use freely for vocal training and music education.